Oliver Elphick wrote:
> Available memory (512M) exceeds the total database size, so sequential
> and random are almost the same for the second and subsequent runs.
>
> Since, in production, I would hope to have all active tables permanently
> in RAM, would there be a case for my using a page cost of 1 on the
> assumption that no disk reads would be needed?
Yes, in your case random_page_cost would be 1 once the data gets into
RAM.
In fact, that is the reason I used only /data/base for testing so places
where data can load into ram will see lower random pages costs.
I could just create a random file and test on that but it isn't the
same.
-- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610)
359-1001+ If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square,
Pennsylvania19073
